Explainable Abstract Trains Dataset

TL;DR

The paper introduces the Explainable Abstract Trains Dataset, a structured image dataset designed to facilitate research on explanation extraction and interpretability in visual classification tasks.

Contribution

It provides a novel dataset with detailed annotations and an ontology for train images, supporting explainability research in computer vision.

Findings

Dataset includes diverse train representations with attribute annotations.

Ontology enables precise classification and understanding of train features.

Supports development of explainability algorithms in visual recognition.

Abstract

The Explainable Abstract Trains Dataset is an image dataset containing simplified representations of trains. It aims to provide a platform for the application and research of algorithms for justification and explanation extraction. The dataset is accompanied by an ontology that conceptualizes and classifies the depicted trains based on their visual characteristics, allowing for a precise understanding of how each train was labeled. Each image in the dataset is annotated with multiple attributes describing the trains' features and with bounding boxes for the train elements.